2003ApJ...598...20D

We investigate the Butcher-Oemler effect using samples of galaxies brighter than observed-frame K^*+1.5 in 33 clusters at 0.1≲z≲0.9. We attempt to duplicate as closely as possible the methodology of Butcher & Oemler. Apart from selecting in the K band, the most important difference is that we use a brightness limit fixed at 1.5 mag below an observed-frame K^* rather than the nominal limit of rest-frame M_V=-20 used by Butcher & Oemler. For an early-type galaxy at z=0.1, our sample cutoff is 0.2 mag brighter than rest-frame M_V=-20, while at z=0.9, our cutoff is 0.9 mag brighter. If the blue galaxies tend to be faint, then the difference in magnitude limits should result in our measuring lower blue fractions. A more minor difference from the Butcher & Oemler methodology is that the area covered by our galaxy samples has a radius of 0.5 or 0.7 Mpc at all redshifts, rather than R₃₀, the radius containing 30% of the cluster population. In practice our field sizes are generally similar to those used by Butcher & Oemler. We find that the fraction of blue galaxies in our K-selected samples is lower on average than that derived from several optically selected samples and that it shows little trend with redshift. However, at the redshifts z<0.6, at which our sample overlaps with that of Butcher & Oemler, the difference in f_Bas determined from our K-selected samples and those of Butcher & Oemler is much reduced. The large scatter in the measured f_B, even in small redshift ranges, in our study indicates that determining the f_Bfor a much larger sample of clusters from K-selected galaxy samples is important.

As a test of our methods, our data allow us to construct optically selected samples down to rest-frame M_V=-20, as used by Butcher & Oemler, for four clusters that are common between our sample and theirs. For these rest-frame V-selected samples, we find similar fractions of blue galaxies to Butcher & Oemler, while the K-selected samples for the same four clusters yield blue fractions that are typically half as large. This comparison indicates that selecting in the K band is the primary difference between our study and previous optically based studies of the Butcher & Oemler effect. Selecting in the observed K band is more nearly a process of selecting galaxies by their mass than is the case for optically selected samples. Our results suggest that the Butcher-Oemler effect is at least partly due to low-mass galaxies whose optical luminosities are boosted. These lower mass galaxies could evolve into the rich dwarf population observed in nearby clusters.